It's pretty much always regulatory stuff. Google is a lawsuit magnet so they are extremely paranoid with this stuff to make sure they're not breaking some obscure local law.
There also could be legal reasons, I have no idea if it's the case here but theoretically I could imagine something like "right to be forgotten" interacting oddly with LLMs.
Also libel (if the LLM makes up facts about someone), copyright (if it reproduces content), privacy (if some of the data it trained on wasn't intended to be public), etc.
And on top of that, there are plenty of novel things that may or may not be illegal. If it can be tricked into making a pro-Nazi statement, is it violating German law? What if it offers medical or legal advice without a license?
It's a pretty big legal minefield, and different jurisdictions have different laws and standards, so it makes sense to limit your exposure.
We don't want that. Every time Google launches something it takes months for it to understand that just because I'm currently located in country X does not mean I want localization or personalization for that country. Just release the service for those that can use it without the localization. After all we know that the vast majority of the models are trained on English, it's useless to give me a subpar undertrained model on local language.
If that's the problem then they're not supplying the right fix. Surely restricting this to users located in the US doesn't prevent people from writing in a language that isn't English. Google should be capable of detecting the language of a query, and potentially rejecting it based on that with an apology.
> PR,
Obviously there must be _some_ reason why they're doing this. Doesn't mean it's a good reason.
Why? They have no idea how many users they will get. If they buy A100s on the assumption they will get 50M daily active users they run the risk of wasting an enormous amount of money if they get 1M users instead. And its not like these GPUs grow on trees. Clearly MSFT is struggling to set up compute fast enough, see the decreasing rate limits on GPT-4.
So you're saying they might be incompetent enough for their estimation to be off by 50x? You're also saying Google, _a cloud computing provider selling access to A100s_, can't scale this dynamically?
> Clearly MSFT is struggling to set up compute fast enough, see the decreasing rate limits on GPT-4.
Well, even so, this is how you'd deal with capacity problems, rather than by arbitrarily shutting out parts of the world.
Predicting the future correctly isn't a matter of competence... And yes, even Google can't scale infinitely in the face of an actual physical resource constraint. When I worked there, there was a period when there was a shortage of memory chips, which required a lot of creativity. I suspect the current period is very constrained by how fast AI/ML focused chips can be manufactured.
> Predicting the future correctly isn't a matter of competence... And yes, even Google can't scale infinitely in the face of an actual physical resource constraint.
In order to not geo-block, Google would need to be able to scale infinitely? What kind of straw man is this?
> there was a shortage of memory chips, which required a lot of creativity.
Geo-blocking doesn't strike me as particularly creative as far as solutions go. Whatever problem they're trying to address, the excuses being made here seem weak and don't make Google look any less incompetent.
:shrug: Yeah it's actually very difficult to launch a very resource intensive product globally all at once. It's not really an excuse, it's just a real problem that Google has. It limits what they can launch and how they can do it, because anything they launch will get used by a huge number of people right away, and everyone expects them to achieve a magical level of performance at that scale. Rolling out region by region based on where data centers are located and what hardware is deployed in them is one possible way to deal with this. I don't think it's good, really, I just think it's unsurprising.
>So you're saying they might be incompetent enough for their estimation to be off by 50x?
Not sure about incompetence, it's just a very hard problem. Sam Altman's estimations for ChatGPT were off by 10x, for reference. Apparently some employees thought it would be a total flop that would barely make the last page of major newspapers.
> You're also saying Google, _a cloud computing provider selling access to A100s_, can't scale this dynamically?
I'm saying that Google can not acquire enormous quantities of A100s overnight, yes.
if you sell something, then sure goods might have laws. but just text ? lol